Advance curve optimizing in practice

[80Stang]

Howdy, I'd like your opinions how this would work and suggestions to get best results out of it.

The whole thinking is based on that you can find the right advance to certain RPM (at certain driving conditions) by finding the peak manifold vacuum at that RPM. Right?

Naturally it is impossible to find it under varying loads but finding it at steady load seems possible and pretty easy.

I have means to control advance from the front seat by independently controlling the vacuum diaphgram of the dizzy. Also I have the meter to read manifold vacuum at all times. I thought I'd have a friend drive the car on a flat road thru a series of RPM levels and me finding the highest manifold vacuum for every RPM level by actuating the vacuum advance on the fly. I need to know (measure)

- initial advance
- centrifugals curve i.e. how much advance it gives thru rpm range
- how much vacuum applied to the diaphgram gives advance

and then it is possible to calculate the advance curve at the "sweet spot" where manifold vacuum was the highest for every RPM level measured. At least that would optimize the curve every time the driving condition is the same; same load, same gear, same everything...

Ok, put your knowledge into play. Would this be a way to a pretty good overall advance curve? Maybe knock appear while the manifold vacuum is still improving, then I'll just mark where the knock appeared and set the curve a pair of degrees lower at that point.

Weather you shoot the idea down or don't, I'll try it anyway. Got the gear, just need to get some other things done first as they didn't give us 28 hours a day.

 

[MustangSix]

I like the approach, but have some questions.

Vacuum may give you a measurement for maximizing VE, but not necessarily for optimal cylinder pressure and power. By optimizing VE by adjusting spark, you may not necessarily optimize power output, which may occur at less than optimal VE levels.

A mechanical advance governed by springs and weights will be able to do no better than generate a two dimensional curve. It will always be dependent on engine speed, so at any speed x, you will only be able to generate advance y. Adding the vacuum gives you a z component to discriminate load, but again, it's a 2D curve governed by a vacuum diaphragm acting against a spring.

Here's the question: How will you make the mechanical advance reduce the advance curve if you find you need less advance at various speeds? What if your best advance is less at 2200 rpm than it is at 1800 but more at 2600? Same with the vacuum advance. What if you find your curves are not smooth, but actually looks like rolling hills?

The problem lies in that you cannot make dynamic adjustment, but your method should get you a fairly good map that works under most conditions.

 

[80Stang]

Thanks for you insight.

Here's the question: How will you make the mechanical advance reduce the advance curve if you find you need less advance at various speeds? What if your best advance is less at 2200 rpm than it is at 1800 but more at 2600? Same with the vacuum advance. What if you find your curves are not smooth, but actually looks like rolling hills?

Can't do. But that's the way engines with mechanical advance have always worked; you can only have more advance while rpms increase. No downhills in between. Then it comes to optimizing, need to find the best average setup to suit the whole rpm range. Unless there are different vacuum diaphgrams at hand or an adjustable one, the whole adjusting needs be done at the centrifugal advance. So I think a adjustable vacuum would be handy as it is much harder to try different setups inside the dizzy.

Computer controlled spark of course would follow the optimal curve no matter what it looks like.

The problem lies in that you cannot make dynamic adjustment, but your method should get you a fairly good map that works under most conditions.

That's the purpose of this research; develop a method to optimize most cars to at least good setup when it comes to advance curves. Once again, it's part of my learning process. Any more thoughts?

 

[MustangSix]

I've just finished testing and setting up the initial ignition and fuel maps for the new Megasquirt II that's going into the coupe. The fuel curves can be dynamically adjusted thru an EGO feedback and the ignition curve can be remapped based on knock sensor readings.

I recall someone saying that in the old days, labor was cheap and technology was expensive. Now technology is cheap and labor is expensive. There is the rub.

In order to make an old tech distributor work well, it takes an infusion of skilled labor, perhaps 8 or 10 hours to pull the device, clean it, tune it, reinstall it, test it, tune it again.... At current rates, that may cost as much as $700 to $1000 if you had to pay for it!

With technology, it may be possible to install an electronic timing control far quicker, for less money, that outperforms the mechanical device. For example, a programable variable curve controller is less than $300 these days.

Anyway, just a thought.

 

[80Stang]

Electronics make it easier to set up the curve, but still, how do you know what kind of curve you need? Vast majority of cars don't have EGOs or knock sensors. We need to know the optimal curve weather the timing was controlled mechanically or electronically.

I'd say it is a fair target using these methods to get the good curve on paper in one hour. To get the curve into the car is another chapter.

 

[Danielson]

I have been thinking of doing exactly the same thing as you mention here and would love to here your results. I have found in my old ford ute that it likes a lot of initial advance (18 degrees), but when setup like that it has too much along with the vacuum. At this stage I have dissconected vac advance and just run mechanical and it is much better. But with the right curve could be better again.

 

[MustangSix]

Electronics make it easier to set up the curve, but still, how do you know what kind of curve you need? Vast majority of cars don't have EGOs or knock sensors. We need to know the optimal curve weather the timing was controlled mechanically or electronically.

I'd say it is a fair target using these methods to get the good curve on paper in one hour. To get the curve into the car is another chapter.

Very true. Once you have the curve determined, you still have to implement it, mechanical, vacuum, or electronic.

Maybe you could combine the three. In those rpm ranges where you may need to retard the curve, you could pull the advance back electronically. Or replace the centrifugal system with an electronic system, but suplement it with a vacuum advance for load sensing.

 

[80Stang]

I'll definitely come back to this when I manage to get the time to do it. Prepared a spreadsheet for this already.